The first two authors contributed equally to this paper.
Infectious Causes of Cancer
Human papillomavirus as a risk factor for the increase in incidence of tonsillar cancer
Version of Record online: 21 SEP 2006
Copyright © 2006 Wiley-Liss, Inc.
International Journal of Cancer
Volume 119, Issue 11, pages 2620–2623, 1 December 2006
How to Cite
Hammarstedt, L., Lindquist, D., Dahlstrand, H., Romanitan, M., (Dahlgren), L. O., Joneberg, J., Creson, N., Lindholm, J., Ye, W., Dalianis, T. and Munck-Wikland, E. (2006), Human papillomavirus as a risk factor for the increase in incidence of tonsillar cancer. Int. J. Cancer, 119: 2620–2623. doi: 10.1002/ijc.22177
- Issue online: 20 OCT 2006
- Version of Record online: 21 SEP 2006
- Manuscript Accepted: 9 JUN 2006
- Manuscript Revised: 2 JUN 2006
- Manuscript Received: 9 MAR 2006
- Swedish Cancer Foundation
- The Stockholm Cancer Society
- The Swedish Research Council
- The Laryngeal Foundation
- Stockholm City Council
- Karolinska Institutet
- human papillomavirus;
- tonsillar cancer;
Smoking and alcohol are well-known etiological factors in tonsillar cancer. However, as in cervical cancer, human papillomavirus (HPV) is currently found in a sizable proportion of tonsillar cancer. Recent reports from the U.S. and Finland show an increase in the incidence of tonsillar cancer, without a parallel rise in smoking and alcohol consumption. This study investigates whether the incidence of tonsillar cancer has also changed in Sweden and whether a possible explanation of the increase is a higher proportion of HPV-positive tonsillar cancer. The incidence of tonsillar cancer between 1970 and 2002 in the Stockholm area was obtained from the Swedish Cancer Registry. In parallel, 203 pretreatment paraffin-embedded tonsillar cancer biopsies taken during 1970–2002 from patients in the Stockholm area were tested for presence of HPV DNA by PCR. The incidence of tonsillar cancer increased 2.8-fold (2.6 in men and 3.5 in women) from 1970 to 2002. During the same period, a significant increase in the proportion of HPV-positive tonsillar cancer cases was observed, as it increased 2.9-fold (p < 0.001). The distribution of HPV-positive cases was 7/30 (23.3%) in the 1970s, 12/42 (29%) in the 1980s, 48/84 (57%) in the 1990s and 32/47 (68%) during 2000–2002. We have demonstrated a highly significant and parallel increase both in the incidence of tonsillar cancer and the proportion of HPV-positive tumors. Hence, HPV may play an important role for the increased incidence of tonsillar cancer. This should definitely influence future preventive strategies as well as treatment for this type of cancer. © 2006 Wiley-Liss, Inc.
Head and neck squamous cell carcinoma (HNSCC) represents a significant cause of morbidity worldwide. It constitutes 3–5% of the malignancies in Europe and in the United States, while in South East Asia and India it reaches up to 40–50%.1, 2, 3 Eighty to ninety percent of head and neck cancer cases are considered to be associated with known risk factors, such as smoking, betel nut chewing and alcohol abuse.3, 4 Nevertheless, 15–20% of head and neck cancer cases occur in individuals without a history of tobacco or alcohol use.5 Accumulating molecular and epidemiological data indicate that high-risk types of human papillomavirus (HPV), similar to those in cervical cancer, are associated with a subset of head and neck cancer (HPV-HNSCC).5, 6, 7, 8, 9 Furthermore, some studies have shown that patients with HPV-HNSCC are less likely to have a history of smoking.8, 10 The strongest association of HPV to HNSCC has been found in oropharyngeal squamous cell carcinoma, especially in the tonsils, where HPV DNA is present in 45–70% of the cases,5, 6, 8, 10, 11, 12, 13, 14 and in base of tongue cancer (40%).15 Moreover, the presence of HPV in tonsillar and base of tongue cancer has been found to be a strong favorable prognostic factor independent of TNM stage, nodal status, gender and age.11, 15, 16
Tonsillar squamous cell carcinoma (SCC) is the most common oropharyngeal tumor site in Sweden17 and in the United States it represents 15–20% of all intraoral and oropharyngeal carcinomas.18 Moreover, the incidence of tonsillar carcinomas has increased 2 and 3% per year from 1973 to 1995 in Caucasian and African–American men, respectively, while the cancer incidence in other oral sites has remained constant.18 Similarly, recent data from Finland also show an increase in the incidence of tonsillar cancer in spite of a decreasing trend in smoking incidence.19 In Sweden, there is also an increase in incidence of tonsillar cancer20 despite a decreasing prevalence in smoking.21
The aim of this study was therefore to investigate whether the incidence of tonsillar cancer had also changed in the Stockholm area between 1970 and 2002, and moreover to investigate whether there was a parallel increase in the proportion of HPV-positive tonsillar cancer within the same area and period of time.
Patients, material and methods
The Swedish Cancer Registry was used to identify all cases of tonsillar cancer from 1970 to 2002 in the metropolitan Stockholm area (with almost 2 million inhabitants). The Swedish Cancer Registry uses the seventh revision of the International Classification of Diseases (ICD-7) and a special code for histological type was also applied. Tonsillar carcinoma, here defined as diagnostic code 145.0 according to ICD-7, included only the squamous cell type.
For tumor biopsy selection, all 515 patients diagnosed with primary tonsillar squamous cell carcinoma between 1970 and 2002 in Stockholm, and reported to the Swedish Cancer Registry, were included in the study, which was conducted with ethical permission 03-386 from the ethics committee of the Karolinska University Hospital. From these 515 patients, 237 diagnostic pre-treatment tumors biopsies were obtainable and included in the study. All specimens were reviewed by a second pathologist for verification of the diagnosis and to ensure that they consisted of at least 70% cancer cells. Specimens containing less than 70% cancer cells were micro-dissected. Hence, samples from approximately half of all patients were not available for further testing, but the distribution of nonavailable samples was similar between the different decades.
After re-evaluation of the histology and primary diagnosis, and loss of samples due to technical problems, 218 cases were included in the study and of these 203 had amplifiable DNA and were analyzed for presence of HPV DNA.
The 203 patients diagnosed with tonsillar cancer who were analyzed for presence of HPV DNA were between 30 and 88 years, with a median age at 62 years. Fifty-eight patients (29%) were females and 145 (71%) were males. Notably, the 312 patients, who were not included in the study, were similar to the 203 included patients, with regard to age (24–91 years of age, median age 62), and gender distribution of 90 (29%) females and 223 (71%) males.
Detection of HPV DNA
DNA was extracted from 25 μm slices from the paraffin-embedded tonsillar cancer biopsies using the High Pure RNA paraffin kit (Roche Diagnostics) according to the manufacturer's instructions, but with exclusion of DNase treatment.
All samples were run in an polymerase chain reaction (PCR) directed to detect the human house keeping gene S14 to verify the presence of amplifiable DNA.22 To further sensitize the S14 detection protocol, a 25 ng sample DNA was run in a total volume of 25 μl with 12.5 μl iQ™ SybrGreen Supermix (BioRad) and 15 pmol of each primer in an iCycler (BioRad). The program consisted of 10 min denaturation at 94°C and 40 cycles of annealing at 57.7°C for 15 sec, elongation at 74°C for 30 sec and denaturation at 94°C for 15 sec. A melt curve starting at 40°C and increasing by 0.5°C every 10th second up to 120°C was run to verify specificity of the obtained amplicons.
Presence of HPV DNA from a broad range of HPV types was analyzed in S14 positive samples by PCR, using the general primer pairs GP5+/6+ and CPI/IIG as previously described.23
To avoid false negative results due to loss of E1 and L1, samples negative in the GP5+/6+ and CPI/IIG runs were also analyzed for HPV-16, which is the most commonly detected HPV type in tonsillar cancer in Sweden.11, 22, 23 HPV-16 was detected in a TaqMan real-time HPV-16-specific PCR directed against the E6 gene. The reaction mixture consisted of 12.5 μl TaqMan® Universal Master Mix (Applied Biosystems, Stockholm, Sweden), 7.5 pmol of each primer, 2.5 pmol probe and 0.5 μl Rox reference dye in a total of 25 μl.
Samples that were negative in the HPV-16 run were sequenced, using direct sequencing of the amplicon from the general PCRs (GP5+/6+ and CPI/IIG) as previously described.23
The Swedish 1970 population was used for direct standardization of tonsillar squamous cell carcinoma incidence rates over calendar years. A Chi-square test was used to compare prevalence rates of HPV DNA in tumor samples across different calendar periods. A Mann–Whitney test was used to compare age in the HPV positive vs. HPV negative patient groups.
Incidence of tonsillar cancer
The age-standardized incidence of tonsillar cancer increased from 1.3 to 3.6 (2.8-fold) per 100,000 between 1970 and 2002 in the Stockholm area (Fig. 1). In men, the increase was 2.6-fold (1.077–2.81 per 100,000), while in women the increase was 3.5-fold (0.232–0.829 per 100,000) (Fig. 1).
Incidence of HPV in tonsillar carcinoma
A total of 99 (49%) of the samples were positive for HPV by either PCR protocol (Table I). Eighty of these samples were positive with the broad spectrum HPV primers, and an additional 19 were detected using HPV-16-specific primers. Eighty-six (87%) of the samples were HPV-16 positive while 12 (12%) were not HPV-16 positive. Direct sequencing of the products from the general PCRs revealed that 3 samples were HPV-33 positive, 1 sample HPV-35 positive and 1 sample HPV-45 positive. The remaining 7 HPV-positive samples were not possible to type verify by sequencing. The distribution of HPV-positive tonsillar cancer biopsies over time was as follows: 7/30 (23%) from the 1970s, 12/42 (29%) from the 1980s, 48/84 (57%) from the 1990s, and 32/47 (68%) from 2000–2002 (Table I). The distribution of samples with amplifiable DNA per number of patient samples over time was 30/39 (77%) from the 1970s, 42/46 (91%) from the 1980s, 84/86 (98%) from the 1990s and 47/47 (100%) from 2000–2002. Hence, there was an increase in the proportion of HPV-positive tonsillar cancer in the period 1970–2002 (Table I). More specifically, there was a significant increase in the proportion of HPV-positive tonsillar cancer 1990–1999 and 2000–2002 compared to 1980–1989 2.0-fold (p = 0.0045) and 2.4-fold (p < 0.001), respectively. Furthermore, the increase in the proportion of HPV-positive tonsillar cancer 1990–1999 and 2000–2002 compared to 1970–1979 was 2.45-fold (p = 0.003) and 2.9-fold (p < 0.001), respectively.
|Calendar years||Number of cases of tonsillar squamous cell carcinoma||Number of pretreatment biopsies retrieved||Excluded1||Presence of HPV DNA2||p-Value3|
Patient features in the HPV positive and HPV negative patients
There was a significant difference in median age between patients with HPV positive and HPV negative tonsillar cancer (p < 0.001). Patients with HPV positive cancer, had a median age of 55 years, 70 patients were males between 30–84 years of age (median age 55 years), while 29 patients were females between 43 and 78 years of age (median age 55 years) (Table II). Patients with HPV negative cancer had a median age of 65, 75 patients were males between 37 and 88 years of age (median age 65 years) and 29 patients were females between 42 and 78 years of age (median age 65 years).
|Patients with HPV positive tumors||Patients with HPV negative tumors|
In this study, we report that the incidence of tonsillar cancer has increased 2.8-fold in the Stockholm area during 1970–2002, in line with the 2.1-fold increase seen in the rest of Sweden.20 Furthermore, from an analysis of available and DNA-amplifiable pretreatment tonsillar cancer biopsies, we have shown that the prevalence of HPV-positive tonsillar cancer in the Stockholm area increased 2.9-fold between1970 and 2002. Moreover, patients with HPV positive tonsillar cancer had a lower median age (55 years), compared to patients with HPV negative tonsillar cancer (65 years).
Hence, in Sweden as in the U.S.18 and in Finland,19 the incidence of tonsillar cancer, and particularly in men, has increased substantially over the past decades, while the incidence of oral cancer, ICD-7 141, 143 and 144, has shown a decrease in Sweden over the last 10-year period. During the same time period, smoking has decreased considerably in Sweden,24 however, in this study, data on smoking and alcohol consumption were not available for individual patients.
In parallel, we found that the proportion of HPV-positive tonsillar cancer had increased significantly over the past 32 years within the Stockholm area. Since around 30% of all patients in Sweden are treated within this area, our study material should be fairly well representative of patients in Sweden as a whole. It was possible to identify amplifiable DNA in samples from the whole period, although the percentage of amplifiable material was somewhat lower (77%) during the 1970s compared to the period later than 1980 (>90% amplifiable samples). However, since the increase in the proportion of HPV-positive tumors was significant not only from the 1970s but also from the 1980s, the increase in the incidence of HPV-positive tonsillar cancer cannot be explained by differences in the ability to amplify DNA from the collected samples.
We suggest that HPV infection within the oropharynx has contributed to the increase in tonsillar cancer incidence in Sweden. This may be the case also in other countries where tonsillar cancer incidence is increasing e.g. in the U.S. and in Finland.
Accumulating evidence supports the existence of two pathways for the pathogenesis of tonsillar cancer: one traditional pathway driven by tobacco and alcohol, and the other by HPV. This could also explain the difference in median age between patients with HPV positive and HPV negative tonsillar cancer. Nevertheless, data also show that combined exposure to HPV-16 and tobacco yields an additional risk for oral cancer beyond that expected with either exposure alone.28
Molecular biological studies indicate that HPV-16 DNA in particular is a risk factor for tonsillar cancer and that the HPV-16 oncogenes E6 and E7 are generally expressed in HPV-positive tonsillar carcinoma.6, 13 Furthermore, antibodies to these oncoproteins have been demonstrated in the sera of HPV HNSCC patients.25 Moreover, in a nested case–control study, in which sera from cases were drawn an average of 9 years before the development of cancer, HPV-16 seropositive individuals had a greater than 14-fold increase in risk of oropharyngeal cancer than seronegative individuals.9
It may be possible that the increase in HPV-positive tonsillar cancer is due to changes in sexual habits, with HPV-16 infections, not uncommon in the urogenital area, also becoming more common in the oropharynx due to an increase in oral-genital sex. The latter is also supported by changes in the location of Herpes simplex viruses (HSV) I and II over time. Thirty years ago HSV I was regularly detected above the waist, while HSV II was generally detected below the waist, while today both HSV I and II are frequently found above the waist.26
One of the few studies on adults with healthy tonsils demonstrated high-risk HPV DNA in 2.2% of the cases. The risk of HPV infection was increased with HIV-seropositivity, HSV-2 seropositivity and sexual behavior defined as >5 lifetime oral sex partners and >5 “casual” sex partners.27 It has also been shown that the risk of oral and oropharyngeal cancer increases with decreasing age of intercourse debut and increasing number of partners in men.28 Furthermore, patients reporting a history of oral-genital sex are more likely to present an HPV-positive oral and oropharyngeal cancer when compared to those without such a history (odds ratio 4.2; 95% CI, 1.5–11.7) after adjustment for gender, tumor location and tobacco use.29
The presence of HPV in tonsillar cancer has favorable impact on prognosis,8, 11 which may relate to both biological and immunological factors yet to be explored. We propose that by examining the HPV status of the tumor, it may be possible in the future to design more specific treatment strategies with the objective to tailor treatment for individual patients. Moreover, ongoing HPV vaccination trials for the prevention of cervical cancer may be effective not only in protecting the cervix from persistent HPV infection and the development of cervical cancer, but also in the prevention of HPV infection in the oropharynx.
In conclusion, we have demonstrated a parallel increase both in the incidence and the proportion of HPV-positive tonsillar cancer. We hypothesize that an “epidemic” of HPV infection in the oropharynx, due to changed sexual habits, may contribute to the significant increase in incidence of tonsillar cancer. We also suggest that the presence or absence of HPV in tonsillar cancer should be considered when tailoring treatment. Finally, the association of HPV to tonsillar cancer may possibly pave the way for the development of prevention strategies, such as vaccination against HPV infection.
The authors thank Erik Onelöv for assistance with the statistics and Juhua Luo for help with the incidence rates.
- 17Vård-program 1999, Huvud-hals-och esofaguscancer. Stockholm: Onkologiskt Centrum Stockholm-Gotland; 1999., , , , ,
- 20The Cancer Registry, Cancer incidence in Sweden 2002, The Swedish Board of Health and Welfare, 2003.
- 21Statistics Sweden. Statistiska Centralbyrån. Hälsa (health) SOS., 2004.
- 26Genital herpes in young adults: changing sexual behaviours, epidemiology and management. Herpes 2005; 12: 10–14..